Ancestor — Case Study

“I wish I had asked more.”

I began by using Claude to scan Reddit conversations for recurring emotional pain points around family memory and loss. One pattern kept resurfacing: people regretted the questions they never asked while older relatives were still alive.

01 · Source

AI-assisted scan of Reddit conversations

I used AI to accelerate the search across many fragmented conversations rather than manually browsing one thread at a time.

02 · Signal

Stories disappear before families realize what they wanted to ask.

The recurring pain was not simply losing photos or recordings. It was losing context, voice, and the chance to ask follow-up questions.

03 · Design question

How might fragmented family memories become something people can actually keep and return to?

Judgment

AI surfaced the signal. Choosing which signal mattered was the first design decision.

The opportunity wasn’t recording. It was curation.

My first concept treated every promising feature as part of the product. AI made it easy to generate breadth; the harder work was deciding what the product should actually stand for.

Initial breadthAI-generated

Guided voice interviewsText-to-memory generationIntelligent photo matchingCinematic playbackFamily sharingStory generation

Useful ideas individually, but together they described a feature set rather than a focused product.

Second lensGemini analysis

I brought the feature set to Gemini as a second product and market-analysis perspective. That reframed the opportunity: recording and content generation already had many existing solutions. The harder problem was what happened after capture.

More memories do not automatically create meaning.

Signal

Families lose stories they wish they had captured.

Tension

Capturing more creates even more photos, audio, and fragments to organize.

Opportunity

Help people turn fragments into a coherent family artifact.

Where the initial ideas landed

3 / 6

of the initial feature ideas ended up serving curation, the job the product was rebuilt around.

Each block is one idea from the initial feature set, grouped by the product job it maps to in the final strategy.

Guided voice interviews

Capture1

Text-to-memory generation

Intelligent photo matching

Cinematic playback

Curate3

Family sharing

Story generation

Preserve2

The shift

Capture more→Curate meaning

My decision

AI contributed breadth and analysis. Making curation the center of the product was my call.

Three jobs shaped the experience.

Once curation became the core idea, I used one simple filter for feature decisions: does this interaction help a memory move from fragment to story?

01 · Capture

Help people remember, not fill out a form.

The experience begins with the person rather than the database: who the story is for, the relationship, and a topic that feels natural to talk about. During recording, the interface stays deliberately sparse so attention remains on the conversation.

Design decisions

Recipient-first setup
AI-guided sensory prompts
One question at a time

Recipient setup

Guided recording

02 · Curate

Surface the moments worth keeping.

The product does not treat every transcript, photo, or quote equally. Recorded material is reorganized into stronger moments, matched with relevant imagery, and shaped into narrative sequences that feel intentional rather than chronological by default.

Design decisions

Semantic highlights
Photo-memory matching
Cinematic memoir generation

Analyzing → cinematic memoir → photo matching

03 · Preserve

Turn memories into an artifact families can return to.

The final output is designed less like cloud storage and more like an heirloom: collected volumes, chapter structures, page turns, bookmarks, and storybook layouts give the archive a sense of permanence.

Design decisions

Collected volumes
Narrative chapters
Digital / physical storybook

Writing → chapters → family book

Making digital memory feel tangible

A family archive shouldn’t feel like a productivity app. I borrowed from objects people already associate with keeping, rereading, and passing stories down.

The Ancestor Archive: a Childhood Memories volume with a mounted photo and an Inspect Volume button — Study 01
Skeuomorphic archive

Recorded memories appear as collected volumes rather than files in a list. The bound-book metaphor gives each story a sense of weight and completion.

Chapter 2 title card, Rosa’s Sunday Bread, with a pull quote over a dark forest photo — Study 02
Editorial storytelling

Serif type, chapter openings, pull quotes, and full-bleed imagery turn generated content into something closer to a memoir than a transcript.

Study 03

Memory as film

Cinematic playback uses photography, pacing, and restrained captions to make a recorded memory feel experienced rather than merely summarized.

Design fingerprint

Linen & paper textureBookmarksPage turnsProgress earned through recordingEditorial imagery

AI expanded the search space. I decided what deserved to survive.

01 · Feature list → archive object

AIHelped generate many ways to organize memories.

My interventionI rejected the productivity-app framing and turned recordings into collected volumes.

02 · Categorized content → narrative pacing

AICould assemble and rewrite memory fragments quickly.

My interventionI reorganized them into chapters, openings, pull quotes, and cinematic sequences.

03 · Generated UI → designed artifact

AIAccelerated prototypes and implementation.

My interventionWhen prompting stopped improving hierarchy, typography, or pacing, I moved into Figma and code and finished the interaction directly.

Judgment, not generation, became the bottleneck.

AI let me explore more directions, faster. The design work was deciding which ones belonged in the product — and knowing when to stop prompting and start making.

“I wish I had asked more.”

AI-assisted scan of Reddit conversations

Stories disappear before families realize what they wanted to ask.

How might fragmented family memories become something people can actually keep and return to?

The opportunity wasn’t recording. It was curation.

Three jobs shaped the experience.

Help people remember, not fill out a form.

Surface the moments worth keeping.

Turn memories into an artifact families can return to.

Making digital memory feel tangible

Skeuomorphic archive

Editorial storytelling

Memory as film

AI expanded the search space. I decided what deserved to survive.